OpenAI, Anthropic, Google and Microsoft (MS), among other major global corporations, urged a joint response, saying cyberattacks leveraging artificial intelligence (AI) are rapidly becoming more sophisticated.
On the 27th (local time), about 120 corporations in the technology, security and finance sectors signed an open letter released primarily by U.S. corporations. In AI and data centers, OpenAI, Anthropic, Google, MS and Amazon Web Services (AWS) signed on; in security, Palo Alto and SentinelOne; and in finance, Visa, Mastercard and U.S. Bank.
In the letter, they said, "In the coming months, as the performance of models around the world continues to improve, AI-driven cyberattacks will become far more widespread and sophisticated," adding, "The time to strengthen cyber defenses is limited."
The letter noted that the damage from AI cyberattacks could spread beyond the internet to corporations and public services that society relies on, such as hospitals and water treatment facilities.
At the same time, they said AI technology, while a means of attack, can also serve as a defensive tool to identify and fix security vulnerabilities accumulated over years, calling for bolstering security capabilities using AI.
They emphasized that rather than confronting cyberattacks individually, a joint response is needed to raise security standards globally and seek solutions.
They also said governments should coordinate cyber defense at the national or international level and directly fund essential public services that lack security budgets.
The letter said attackers must be made to pay a price, and it also called for stronger sanctions and penalties against malicious hackers.
In particular, it urged corporations developing state-of-the-art AI models to provide model access, training, and financial and technical support to essential infrastructure operators and similar entities that lack security personnel and resources. It also said mechanisms to track and monitor the activities of AI agents should be established.
Major global corporations are calling for a joint response because a sense of crisis is growing that AI-driven cyber threats are becoming a reality.
Recently, concerns that AI agents could engage in cyberattacks without human involvement have also materialized. A series of related security incidents has been disclosed, including a case in which OpenAI's AI agent, during testing, left an isolated environment and accessed the external organization Hugging Face without authorization.
However, the open letter did not mention security-related support measures or their scale from major AI developers.